linux-kernel - SCM_CREDENTIALS and PID namespaces

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Date:	Mon, 6 Jul 2009 16:13:48 -0700
From:	Adam Langley <agl@...erialviolet.org>
To:	linux-kernel@...r.kernel.org
Subject: SCM_CREDENTIALS and PID namespaces

Currently, the SCM_CREDENTIALS control message contains the PID of the
sending process, in the sender's PID namespace. This would seem to
violate the spirit of SCM_CREDENTIALS since, from the receiver's point
of view, the sender's PID in that namespace might be another process
entirely.

I started to write a patch for this, but then got hung up on the
semantics, so I'm asking here first.

Here's what I think should happen:

A received SCM_CREDENTIALS should contain the PID of the sending
process, in the receiver's namespace. Or -1 if the PID is not
representable.

If the sending process has exited, the pid should be -1. (We don't
want to hold a reference to a struct pid from the SKB, so we have to
do this).

When sending an SCM_CREDENTIALS message, if pid == getpid(), then the
PID acts as above. Otherwise, we pass the PID raw to the receiver.  (A
process has to be CAP_SYS_ADMIN to fake its PID).

Seem reasonable?

AGL

-- 
Adam Langley agl@...erialviolet.org http://www.imperialviolet.org
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/